This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!

(Paraphrased)
The train! statement in the README.md example runs in rougly 7.5 minutes on my machine. I forgot that steps=300 by default.

• Add a @progress macro from ProgressLogging.jl to the training loop in train!.
• Add training loop tracker via callback (cb) as a kwarg to train!.

Originally posted by @elisno in #1 (comment)

I realize the package is at its early days :) , but do you, @VarLad, have a plan for its roadmap?

Would this be along the lines of https://github.com/QuantumBFS/QuDiffEq.jl ?

If yes, I think the references in https://quantumalgorithmzoo.org/#ONML might be a good start :D

Investigation required

This issue has been created to discuss the license of this library further going on.
No new PRs will be made by the author of this issue, before this issue is resolved

cc: @ChrisRackauckas

@JuliaRegistrator register()

Question❓
I have run the test case in the repository. I have a question what is the meaning of M? Does it be a simulation duration? because the plot uses new_M for the x-axis. In addition, if M is the simulation duration when I tried to change M from 0.9 to 10 I got the error that DomainError with 1.001001001001001: asin(x) is not defined for |x| > 1.. If M is not the simulation duration, then how could I generate the solution after train and plot it?

Them you for your time.

I tried to solve the example of the package as follows:

using DifferentialEquations, Yao, QuantumNLDiffEq
# Making the ODEProblem
function f(u, p, t)
	λ, κ = p
	return -1*λ*u*(κ + tan(λ*t))
end
prob = ODEProblem(f, [1.0], (0.0, 0.9), [8.0, 0.1])
function loss_func(a, b)
	return (a - b)^2
end

#Making the DQC
DQC = [QuantumNLDiffEq.DQCType(afm = QuantumNLDiffEq.ChebyshevTower(2), fm = chain(6, [put(i=>Ry(0)) for i in 1:6]), cost = [Add([put(6, i=>Z) for i in 1:6])], var = dispatch(EasyBuild.variational_circuit(6,5), :random), N = 6)]
config = DQCConfig(abh = QuantumNLDiffEq.Floating(), loss = loss_func)
M = range(start=0; stop=0.9, length=21)
evalue(M) = [QuantumNLDiffEq.calculate_evalue(DQC[1], DQC[1].cost, prob.u0[1], config.abh, params[1], M[x], M[1]) for x in 1:length(M)]
params = [Yao.parameters(DQC[1].var)]

#Training the circuit
QuantumNLDiffEq.train!(DQC, prob, config, M, params, steps = 1000)

I got the error as follows:

MethodError: no method matching +(::NamedTuple{(:afm, :fm, :cost, :var, :N, :evol), Tuple{NamedTuple{(:pc,), Tuple{Float64}}, Vararg{Nothing, 5}}}, ::Base.RefValue{Any})

Closest candidates are:
  +(::Any, ::Any, ::Basic, ::Any...)
   @ SymEngine C:\Users\htran\.julia\packages\SymEngine\hozAR\src\mathops.jl:51
  +(::Any, ::Any, ::Any, ::Any...)
   @ Base operators.jl:578
  +(::Union{InitialValues.NonspecificInitialValue, InitialValues.SpecificInitialValue{typeof(+)}}, ::Any)
   @ InitialValues C:\Users\htran\.julia\packages\InitialValues\OWP8V\src\InitialValues.jl:154
  ...

This error happened even though I tried with Julia 1.8, 1.9, or 1.10 for ChebyshevTower and ChebyshevSparse and did not happen for Product
Please help me on how to fix it.

I cannot reproduce this plot from the current example.
I can reproduce the plot from the MWE after updating Yao.jl (see #3 (comment))

It would be nice to make a test case that reproduces the plot, or rather the resulting expectation values.

MWE

julia> begin
           using Yao, QuantumNLDiffEq
           M = range(start=0; stop=0.9, length=20)
           cost = sum([put(6, i=>Z) for i in 1:6])
           var = dispatch(EasyBuild.variational_circuit(6,5), :random)
           quantum_feature_map_circuit = chain(6, [put(i=>Ry(0)) for i in 1:6])
           params = parameters(var)
           u_0 = 1.0
           N = 6
       end
       
julia> train!(quantum_feature_map_circuit, cost, var, M, N, 1.0, params)

julia> e = [expect(cost, zero_state(N)=>QuantumNLDiffEq.new_circuit(quantum_feature_map_circuit, var, i, params, N; mapping=QuantumNLDiffEq.Chebyshev())) .+ u_0 .- expect(cost, zero_state(N)=>QuantumNLDiffEq.new_circuit(quantum_feature_map_circuit, var, M[1], params, N; mapping=QuantumNLDiffEq.Chebyshev()))  for i in M]

julia> import Plots

julia> Plots.plot(M, real.(e))

Expected plot

Actual plot

Originally posted by @elisno in #1 (comment)